paper-with-me

홈 › Papers

DP-AdamW: Investigating Decoupled Weight Decay and Bias Correction in Private Deep Learning

2025-11-11 · Jay Chooi, Kevin Cong, Russell Li, Lillian Sun arxiv

As deep learning methods increasingly utilize sensitive data on a widespread scale, differential privacy (DP) offers formal guarantees to protect against information leakage during model training. A significant challenge remains in implementing DP optimizers that retain strong performance while preserving privacy. Recent advances introduced ever more efficient optimizers, with AdamW being a popular choice for training deep learning models because of strong empirical performance. We study \emph{DP-AdamW} and introduce \emph{DP-AdamW-BC}, a differentially private variant of the AdamW optimizer with DP bias correction for the second moment estimator. We start by showing theoretical results for privacy and convergence guarantees of DP-AdamW and DP-AdamW-BC. Then, we empirically analyze the behavior of both optimizers across multiple privacy budgets ($ε= 1, 3, 7$). We find that DP-AdamW outperforms existing state-of-the-art differentially private optimizers like DP-SGD, DP-Adam, and DP-AdamBC, scoring over 15\% higher on text classification, up to 5\% higher on image classification, and consistently 1\% higher on graph node classification. Moreover, we empirically show that incorporating bias correction in DP-AdamW (DP-AdamW-BC) consistently decreases accuracy, in contrast to the improvement of DP-AdamBC improvement over DP-Adam.

📄 PDF Abstract BibTeX arXiv:2511.07843

Code (0)

등록된 구현이 없습니다.

Tasks

Image ClassificationNode ClassificationText Classification

Similar Papers 제목 키워드 기반

Weight Norm Control

2023-11-19 · Ilya Loshchilov

We note that decoupled weight decay regularization is a particular case of weight norm control where the target norm of weights is set to 0. Any optimization method (e.g., Adam) which uses decoupled weight decay regulari…

Implicit Bias of AdamW: $\ell_\infty$ Norm Constrained Optimization

2024-04-05 · Shuo Xie, Zhiyuan Li

Adam with decoupled weight decay, also known as AdamW, is widely acclaimed for its superior performance in language modeling tasks, surpassing Adam with $\ell_2$ regularization in terms of generalization and optimization…

Language ModelingLanguage Modelling

AdamHD: Decoupled Huber Decay Regularization for Language Model Pre-Training

2025-11-18 · Fu-Ming Guo, Yingfang Fan arxiv

Adaptive optimizers with decoupled weight decay, such as AdamW, are the de facto standard for pre-training large transformer-based generative models. Yet the quadratic nature of the $\ell_2$ penalty embedded in weight de…

Stable Weight Decay Regularization

2020-09-28 · Zeke Xie, Issei Sato, Masashi Sugiyama

Weight decay is a popular regularization technique for training of deep neural networks. Modern deep learning libraries mainly use $L_{2}$ regularization as the default implementation of weight decay. \citet{loshchilov20…

Adam-family Methods with Decoupled Weight Decay in Deep Learning

2023-10-13 · Kuangyu Ding, Nachuan Xiao, Kim-Chuan Toh

In this paper, we investigate the convergence properties of a wide class of Adam-family methods for minimizing quadratically regularized nonsmooth nonconvex optimization problems, especially in the context of training no…

Deep Learning